datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
trellis500k-sketchfab-archivessplataverse-sketchfab
Splataverse: Sketchfab
sketch2stl-parts-tabular
sketch2stl-parts-tabular
A small, hand-measured tabular dataset of 34 desk-scale physical objects, plus 400 label-preserving
synthetic rows, built to support a 2D-sketch-to-3D-print system.
1. Purpose
This dataset was built for Homework 1 of Designing with AI, in direct support of my Project 1 system:
a Gradio application that turns a simple 2D sketch into an editable, 3D-printable STL.
The one decision that determines whether that pipeline produces a correct… See the full description on the dataset page: https://huggingface.co/datasets/sunkaiwen/sketch2stl-parts-tabular.sketch2stl-part-requests-text
sketch2stl-part-requests-text
100 hand-written natural-language requests for 3D-printable parts, each labelled with whether a single
straight extrusion can make the part, plus 1,200 label-preserving synthetic variants.
1. Purpose
Built for Homework 1 of Designing with AI, in support of my Project 1 system: a Gradio app that turns a 2D
sketch into an editable, 3D-printable STL.
My GUI has a free-text box next to the sketch canvas. That text is the earliest signal… See the full description on the dataset page: https://huggingface.co/datasets/sunkaiwen/sketch2stl-part-requests-text.openprocessing-sketches
OpenProcessing Creative Code Dataset
A curated collection of 11642 public creative coding sketches from
OpenProcessing, with full source code.
Sources
Sketches were discovered via:
Curations: [78544, 88104]
User portfolios: [128718, 110137, 363571, 183691, 410675, 314603, 18537, 67809, 244319, 281109, 327947, 314640, 162823, 326951, 387225, 471647, 327946, 411850, 231571, 468480, 211463]
User hearts: [128718, 110137, 363571, 183691, 410675, 314603, 18537, 67809, 244319… See the full description on the dataset page: https://huggingface.co/datasets/t14n/openprocessing-sketches.editable-sketch-repro-results-final
Scaled FormalMath-Lite reproduction for paper mI3K0e1KsN
This repository contains the machine-verification output from an independent,
scaled reproduction of Editable Proof Sketch for Automated Theorem Proving.
It is not a benchmark-scale estimate of the paper's reported 76.0% pass rate.
Valid run
Benchmark: SphereLab/FormalMATH-Lite (8 deterministic examples out of 425,
seed 2022)
Generator: deepseek-ai/DeepSeek-Prover-V2-7B
Verifier: Lean 4.11.0 with Mathlib… See the full description on the dataset page: https://huggingface.co/datasets/Sugutt/editable-sketch-repro-results-final.Cities_Wikipedia_Informationeditable-sketch-repro-resultsNew_City_Dataeditable-sketch-repro-canarySketchVCL